TW397958B

TW397958B - Image processing system

Info

Publication number: TW397958B
Application number: TW088104146A
Authority: TW
Inventors: Seiji Tateyama
Original assignee: Hudson Soft Co Ltd
Priority date: 1992-10-09
Filing date: 1993-09-08
Publication date: 2000-07-11
Also published as: CA2106441C; EP0592120A3; US6208333B1; DE69329671D1; EP0944012A1; DE69329671T2; EP0592120A2; US5515077A; US6380944B2; US5812119A; US20010017613A1; EP0592120B1; TW371340B; CA2106441A1

Abstract

Image data are defined by color vectors each composed of vector factors. Each vector factor specifies the brightness and color difference of a dot. The image data are transmitted with a normal timing when the vector factors correspond to dots of the image one for one, and are transmitted with another timing in another case to scroll the image horizontally.

Description

A7 __ B7 五'發明説明（1 ) 本發明係關於處理自然畫’動畫等之圖像處理系統。以往之技術將圖像朝向畫面左右方向移動之手法叫做水平捲動，上下方向移動之手法叫做垂直捲動。在電動玩具（電腦遊戲機）般之圖像處理裝置中，水平捲動對應於Η遮沒期間，垂直捲動對應於V遮沒期間。圖1乃表示Η遮沒及V遮沒，在此期間並無圖像顯示》因爲是掃描線返回之期間，故把Η遮沒稱爲水平返馳期間，V遮沒稱爲垂直返馳期間。對顯示資料之加工係在此期間進行，如配合顯示期間傳送圖像資料，即可把圖像顯示於電視畫面上。圖像資料係以點（像素）單位的色來表示。色資料有以 3原色（RGB)表示，及以亮度（Υ)和色差（UV)表示之方法。無論那一種，可把這些色依水平方向點單位表成：此一連串色資料稱爲色向量，各點單位之色資料CO，C1， C2....等即稱爲色向量因子。此1向量份即成爲1光柵的顯示。圖2表示各向量因子與點之關係。於此場合，因向量因子與點爲1比1的對應，故如在 1H遮沒前從虛擬畫面（VRAM)讀取圖像資料，朝向其次之水平顯示期間傳送資料，即可顯示1光柵份之圖像。顯示資料儲存於虛擬畫面（VRAM )。虛擬畫面較實畫面 (電視畫面）爲大，其一部份被顯示於實畫面上。虛擬畫面本紙张尺度適川中围國家標準（CNS } A4規格（210X297公釐） (誚先閱讀背面之注意事項再蛾涔本页〕A7 __ B7 Five 'invention description (1) The present invention relates to an image processing system for processing natural paintings' animation and the like. In the prior art, the method of moving the image to the left and right of the screen is called horizontal scrolling, and the method of moving up and down is called vertical scrolling. In an image processing device like an electric toy (a computer game machine), horizontal scrolling corresponds to a blackout period, and vertical scrolling corresponds to a V blanking period. Figure 1 shows the occlusion and V occlusion, and no image is displayed during this period. "Because it is the scanning line return period, the occlusion is called the horizontal flyback period, and the V occlusion is called the vertical flyback period. . The processing of the display data is performed during this period. If the image data is transmitted during the display period, the image can be displayed on the TV screen. The image data is expressed in the color of dots (pixels). Color data are represented by 3 primary colors (RGB), and brightness (Υ) and color difference (UV). Regardless of which type, these colors can be expressed in units of horizontal points: This series of color data is called a color vector, and the color data of each point unit CO, C1, C2, etc. is called a color vector factor. This 1 vector will become a 1 raster display. Figure 2 shows the relationship between each vector factor and a point. In this case, because the vector factor corresponds to a point of 1: 1, if you read the image data from the virtual screen (VRAM) before 1H masking, and transfer the data toward the next horizontal display period, you can display 1 raster copy. Of images. The display data is stored in a virtual screen (VRAM). The virtual picture is larger than the real picture (TV picture), and a part of it is displayed on the real picture. Virtual screen This paper is in accordance with Chuanwei National Standard (CNS) A4 size (210X297mm) (诮 Please read the precautions on the back first, then this page)

*1T 線 A7 B7 —. —. —--—,-- , _ _.„ _ 五、發明説明（2 ) 之那一部份顯示於實畫面，係依圖3所示之BXR及BYR而定。BXR爲點單位之値，BYR爲光柵單位之値。指定BXR 的暫存器就是BGX暫存器，指定BYR之暫存器就是BGY暫存器。* 1T line A7 B7 —. —. —--—,-, _ _. „_ V. The part of the description of the invention (2) is shown on the real screen, which is based on the BXR and BYR shown in Figure 3. BXR is the unit of point unit, BYR is the unit of raster unit. The register of BXR is BGX register, and the register of BYR is BGY register.

例如，欲向畫面右方水平捲動X點之量時，假定現在水平方向之坐標値爲BGX，則把 BGR-X 之値設定於BXR暫存器就可。即，如圖4所示，把自現在顯示者朝向左方向移X個點之量的坐標値做爲顯示的開始點，以傳送資料，則可看到圖像好像向右捲動X個點一般。由於BXR暫存器係從其次之光柵起才有效，故從虛擬畫面讀取圖像資料之動作係於資料傳送的1光柵前進行。如上述，1向量因子對應於1個點，故通常不需經過加工，照原送出就可。但是，若1向量因子非對應於1個點之情況時，端賴BXR暫存器則無法做到。此乃因指定BXR 暫存器之BGX是點單位之故。關於這一點，現在舉出也處理自然畫之圖像處理系統加以說明。以往之電玩裝置係以動畫爲主，且圖像資料的顯示色只以4色，16色’ 256色等少數目的彩色爲己足。因之，如圖5所示，彩色資料只要能以1點單位指出調色盤之位址即可。然而，如要處理自然畫，則必須有幾萬，幾百萬色的顯示色。如使用以往的調色盤，則調色盤變成龐大，會壓迫到記憶體，故並非可行。本紙乐尺度適扣中囤國家標率（CNS ) Α4规格（210X297公釐} (誚先閱讀背面之注意事項再蛾寫本I)For example, if you want to scroll the amount of X points horizontally to the right of the screen, assuming that the current horizontal coordinate 値 is BGX, then you can set the GR of BGR-X to the BXR register. That is, as shown in FIG. 4, the coordinate 値 which is shifted by X points from the present display to the left direction is used as the starting point of the display to transmit data, and you can see that the image seems to scroll X points to the right general. Since the BXR register is effective from the next raster, the action of reading image data from the virtual screen is performed before the 1 raster of data transmission. As mentioned above, a 1-vector factor corresponds to 1 point, so it usually does not need to be processed, and it can be sent as it is. However, if the 1-vector factor does not correspond to 1 point, it cannot be achieved by the BXR register. This is because the BGX of the designated BXR register is a point unit. In this regard, an image processing system that also processes natural painting will now be described. In the past, video game devices were mainly based on animation, and the display color of image data was limited to 4 colors, 16 colors ’256 colors, and a few other colors. Therefore, as shown in FIG. 5, the color data need only indicate the address of the color palette in units of 1 point. However, if you want to deal with natural painting, you must have tens of millions, millions of colors of display colors. If a conventional color palette is used, the color palette becomes large and it will stress the memory, so it is not feasible. The paper scale is suitable for national standard rate (CNS) Α4 specifications (210X297 mm) (诮 Please read the precautions on the back before writing the moth I)

*1T 線- Β 飪浐部中夾^.^1-而hi,消於合竹11印五、發明説明（3 ) 也許可考慮把彩色資訊直接附予與點對應之色資料，俾能行顯示之方法。但是，直接具有彩色資訊時，由於色資料過大，也一樣壓迫記憶體。因此，利用自然畫之相鄰彩色相似的事實，採用把1 色資料（圖2中所述之色向量因子）對應於多個點等之方法。圖6爲可處理自然畫之圖像處理系統之例的系統方塊圖。此裝置之控制單元，圖像資料擴展單元，VDP單元及視頻編碼單元，分別係由1C晶片構成。 KRAM，RRAM，VRAM分別爲控制單元，圖像資料擴展單元，VPP單元之外部記憶體。控制單元係從CD-ROM 讀取資料，先儲存於KRAM後，依需要施以加工，並把壓縮資料轉送至围像資料擴展資單元。圖像資料擴展單元把來自控制單元之壓縮資料解碼，轉送至視頻編碼單元。視頻編碼單元把來自控制單元之資料直接顯示於電視畫面，或把來自圖像資料擴展單元，VDP 單元之圖像資料顯示於電視畫面》現在說明圖像資料擴展單元之機能。圖7爲圖像資料擴展單元之方塊圖。資料匯流排緩衝器爲接受來自控制單元等裝置之圖像資料的記憶體，此記憶體把資料傳送分給預定之區塊。外部記憶（R-RAM)爲儲存解碼資料之區域，其記憶尺寸爲16光柵份之容量（64K位元）x2。爲了提高處理速度，具有2個區域，以行交替讀寫》 -6- 本紙张尺度述州屮囤國家楳準（CNS ) A4規格（2丨0X297公釐） (誚先閱讀背面之注意事項再硪艿本页) ,ιτ 線輕浐部中夾(rif而h-1消炝合竹.！；卬*''4 A7 _________B7_ 五、發明説明（4 ) 在圖像資料擴展單元，可處理IDCT圖像及行進長度 (Run Length)圖像。IDCT圖像爲以IDCT解碼之自然畫、動畫像，行進長度圖像爲受到行住進長度壓縮過之人工動畫像。兩者之壓縮對象圖像均爲每1場256點x 240光栅。 IDCT之顯示色爲約1677萬色，行進長度圖像之行進長度模式爲調色盤方式，有16色/32色/64色/128色之4 種。端子KR0-KR7爲把資料從控制單元傳輸到圖像資料擴展單元之資料匯流排。-REQR爲從圖像資擴展單元傳輸給控制單元之資料請求信號用端子。資料請求信號向控制單元要求壓縮圖像資料。即： -REQR = 0 ____要求資料 -REQR = 1 ____停止資料圖像資料擴展單元於16光柵期間把16光柵份之壓圖像資料解碼。因之，本發明之傳送開始時序之定法係將欲顯示1光柵份之圖像資料自16光柵前起開始向圖像資料擴展單元傳送，並在顯示開始之1光柵以前結束。圖像資料擴展單元因不具光柵資訊，故必須在控制單元決定時序。在圖像資料擴展單元，即與控制單元之資料傳送開始後之第16個HSYNC同步施行圖像之顯示。圖8爲圖像資料之流動及其時序的示意圖。圖中，當電視畫面上正在寫出第3個光柵時，控制單元已把第4個的16光柵份資料傳往圖像資料擴展單元。此傳送處理在本紙張尺度適川中囷囤家標率（CNS ) A4規格（210X297公釐） (誚先閱讀背面之注意事項再"'巧本頁)* 1T line-Β clip in the cooking section ^. ^ 1- And hi, disappeared in Hezhu 11 Seal 5. Invention description (3) It is also possible to consider color information directly attached to the color information corresponding to the point Display method. However, when the color information is directly available, the color data is too large, which also compresses the memory. Therefore, to take advantage of the fact that adjacent colors of natural painting are similar, a method is adopted in which one color data (color vector factor described in Fig. 2) corresponds to multiple points and the like. Fig. 6 is a system block diagram of an example of an image processing system capable of processing natural pictures. The control unit, image data expansion unit, VDP unit and video coding unit of this device are all composed of 1C chip. KRAM, RRAM, and VRAM are the external memory of the control unit, image data expansion unit, and VPP unit, respectively. The control unit reads data from the CD-ROM, stores it in KRAM first, processes it as needed, and transfers the compressed data to the image data expansion unit. The image data expansion unit decodes the compressed data from the control unit and transfers it to the video encoding unit. The video encoding unit displays the data from the control unit directly on the TV screen, or displays the image data from the image data extension unit and VDP unit on the TV screen. The function of the image data extension unit will now be described. Figure 7 is a block diagram of the image data expansion unit. The data bus buffer is a memory that accepts image data from devices such as the control unit, and this memory divides the data transfer into predetermined blocks. External memory (R-RAM) is the area for storing decoded data. Its memory size is 16 raster copies (64K bits) x2. In order to improve the processing speed, it has 2 areas, which can be read and written alternately by line. -6- This paper standard states the state standard (CNS) A4 specification (2 丨 0X297mm) (诮 Read the precautions on the back before reading硪艿 This page), ιτ line light 浐 middle clip (rif and h-1 eliminate the combination of bamboo.!; 卬 * '' 4 A7 _________B7_ V. Description of the invention (4) In the image data expansion unit, IDCT can be processed Images and Run Length images. IDCT images are natural pictures and moving images decoded by IDCT, and running length images are artificial animation images compressed by the running length. The compression target images of both Both are 256 dots x 240 rasters per field. The display color of IDCT is approximately 16.77 million colors, and the travel length mode of the travel length image is the color palette method. There are 4 types of 16 colors, 32 colors, 64 colors, and 128 colors. Terminals KR0-KR7 are data buses for transmitting data from the control unit to the image data extension unit. -REQR is a terminal for data request signals transmitted from the image data extension unit to the control unit. The data request signals are requested from the control unit Compressed image data. Ie: -REQR = 0 ____requested data -REQR = 1 ____stop data The image data expansion unit decodes 16 raster copies of the image data during the 16 raster period. Therefore, the method of the transmission start timing of the present invention is to display the image data of 1 raster copy from the 16 raster From the beginning, it starts to transmit to the image data expansion unit and ends before the raster of the display start. Since the image data expansion unit does not have raster information, it must determine the timing in the control unit. The 16th HSYNC synchronously executes the image display after the data transmission starts. Figure 8 is a schematic diagram of the flow of image data and its timing. In the figure, when the third raster is being written on the TV screen, the control unit has Transfer the 4th 16 raster data to the image data expansion unit. This transmission process is suitable for the standard Chuan Chuan standard (CNS) A4 specification (210X297 mm) (诮 Please read the precautions on the back before " 'This page)

，1T Α7 Β7 五、發明説明（5 ) 電視畫面把第3個之16光柵份顯示完畢以前結束。如此依序處理光柵，即可把一畫面份之資料顯示於電視畫面。此處理稱爲正常再生。圖像資料擴展單元具有來自控制單元之圖像資料用之 FIFO，當 FIFO 告滿時，把失效"DISABLE" ( - REQR = 1)之請求信號送給控制單。控制單元接到此信號時，資料傳送則會暫停。控制單元內如圖9所示具有圖像資料擴展單元傳送控制暫存器，圖像資料擴展單元資料開始址暫存器及傳送區塊數暫存器。圖像資料擴展單元傳送控制暫存器爲用以允許圖像資料擴展單元資料之傳送者，如未獲此允許•則不做傳送。如於傳送中成爲不允許狀態，則在此時機停止傳送。開始位址暫存器係用來指定儲存有像資料擴展單元用資料之KRAM的起頭位址者，控制單元係從KRAM之位址開始傳送。在區塊傳送中，自動累加位址。 #ί浐部中央ir.srxih-T消於合竹ίι印y (ti先間讀背面之注意事項再填寫本Η ) 傳送開始光栅暫存器係用來指定控制單元對圖像資料擴展單元開始資料傳送的時序之暫存器，可指定光柵號數。傳送區塊數暫存器係用來指定以向圖像資料擴展單元傳送之16光柵份資料爲1區塊時之傳送區塊數。如無改變內容，各暫存器在接下來的同一幀中也顯示相同的圖像。各暫存器在設定後即有效，但於正在向圖像資料擴展單元傳送區塊時，即在其傳送後才有效β Ψ n n* ϋ^— 資料之實際傳送，僅限於Κ匯流排K-BSU)被調整》本紙张尺度適川肀囤國家標準（ CNS ) A4規格（210X297公釐）紂浐部中决"4,-^力工消贤-?竹右印" A7 B7 五、發明説明（6 ) 圖像資料擴展單元資料之存取爲可能，且圖像資料擴展有輸出請求信號（-REQR = 0)之情況下進行。資料傳送變成失效之情況，發生於如下之場合。 (1) 圖像資料擴展單元未開始處理之場合。 (2) 圖像資料擴展單元之FIFO告滿之場合。 (3) 自16線之最初資料來到時起計數HSYNC 16次之期間，把埋於資料中之位元組數之所有資料讀入完畢之場合。如欲停止控制單元之傳送，則令傳送開始位元成爲失效。如上所說明，此系統具有16色，32色，64色，128 色，16M色之5種彩色模式，並以YUV方式顯示彩色。Y 表示亮度，U、V表示色差。其中之16色，32色，64色，128色乃如以往一樣使用調色盤，16M色即直接指定彩色資料。其理由是爲了節省記憶體尺寸。舉例而言，如要顯示16色，僅需4位元就夠，但如以此4位元直接當成色資料，則會被固定爲16色。所以準備可從65 536色中指定512個份色資料之調色盤，以便能依每1個畫面指定調色盤之位移，及以上述4位元選擇 1畫面內的調色盤。如此一來，可從65536色中選擇顯示任意的16色，如此可產生柔軟性。然而，如前所述，爲準備16M色用調色盤，需要幾Μ 位元組的記憶體，致難合實用。爲何在16Μ色時要彩色資 -9- 本紙张尺度述扣中囤囤家棉準（CNS ) Α4規格（2丨0X297公釐) 一 (誚先閱讀背面之注意事項再•本頁) ,?τ Μ^‘部屮次榀^-^h.T.消价合竹ίι卬1'! A7 _________B7____ 五、發明説明（7 ) 料直接攜帶彩色資訊之理由在此》此時之問題在於，表示1個點究竟需要攜帶多少位元之色資料。以數量最多的16M色而言，需要有能表示16M 種彩色的大小，因之表示1個點之彩色資訊，則需要24 位元β因爲 16Μ = 16 X IK xlK = 24 χ ΙΟ10 x 1〇10 = 224 是故，把8位元之色資料分別分配給Y，U，V，以表現彩色。KRAM內之虛擬畫面要準備較實畫面爲大者。因其尺寸爲1024 χ 1024點，故如要準備16M色模式之虛擬畫面，即使用單純計算也需要24M位元之記憶體》但由於使用16M色模式者一般都是錄影機或影像掃瞄器所取入之自然畫像，故只需實畫面份之大小即可。換言之，虛擬畫面之1/16程度就可。雖然如此，也需要5M位元。於16M色模式時，爲使色資料每1個點平均具有16 位元，則表成如下： (Y〇 Y, U〇 V〇, Y2 Y3 U, Vl5 ... Υη.2 Υη., Um VJ 式中 m = (η-1)/2 相鄰點之U，V爲共同，差異係以亮度Υ表示。如圖 1 0所示，Υβ Y, uD V。可分爲YQ U。▽。及UQ V。。持有如此圖像資料仍不發生問題，其原因乃爲基於一事實，即在自然畫像中相鄰點之彩色的相差不大。依照上述方式，則能以32位元之色資料表示2個點，每個點平均具有16位元之色資料即可，而非24位元，1 •10- 本紙張尺度通ϋ國國家榡準(CNS ) A4規格（2丨OX297公楚) (誚先閲讀背面之注意事項再iAK-本頁) -β A 7 __________ B7 五、發明説明（8 ) 畫面份之資料僅需1M位元就夠。由此可知，在16M色模式時，1向量因子對應於2點。此向量因子係以1 6光柵單位從控制單元送到圖像資料擴展單元。於1對1的向量因子之情況時，因將所讀取之向量因子原封不同送出即可，故容易控制。但是於1對2之向量因子之情況，則不然。因爲奇數點之水平捲動時，也會把非顯示之點送來之故。需要向量因子之讀取時序的理由在此》需要再次說明者，就是視頻畫面係由像素（點）之集合構成，把這些像素發色，可作出種種圖像。此時依每個點使用調色盤之色》在電腦之處理中，如圖1 1所示，係藉調色盤碼指定於與點對應之色資料之方式實現之。調色盤內之彩色中，可同時使用之彩色之數目稱爲色模式。例如同時使用16色時，則稱爲16色模式。通常電腦遊戲機使用之電視畫面爲256 X 240點。亦即1畫面份有 61，440 點。 16色模式時，色資料之尺寸需要4位元，故1畫面份之色資料之尺寸爲： 61440 X 4 =245760位元=30720位元組=30K位元組色模式之數目愈多，色資料之尺寸要愈大，總資料量也變成龐大。如要於短時間內傳送此種資料量，耗時多，對於需要短時間處理圖像之電玩機造成大的阻礙。因此，如圖1 2所示於傳送資料時，把資料壓縮（碼化），另在接 -11· 本紙张尺度璉瓦不闼國家榇準（CNS ) Α4規格（HOX297公蝥) (誚先間讀背面之注意事項再蛾朽本頁) 訂 A7 B7 五、發明説明（9 ) 受此資料之一方把此資料以解碼裝置解碼後予以圖像顯示。因一般圖像之相鄰點之色資料相同之情形多，故圖像資料係以 <調色盤碼> + <個數> 之型式壓縮。圖13係表示壓縮資料與點之關係。解碼時，藉調色盤碼（PLT)A之個數被重複，可回復原來的圖像資料。此個數稱爲調色盤表（CRL)。因之，此種壓縮法叫做行進長度法（Run Length Method)。依照以往，色模式只能在1畫面使用1模式。因之，儘管<PLT> + <CRL>爲固定長度，亦不致發生問題。但若爲1畫面可指定多個色模式時，固定長度之<PLT> + <CRL> 會產生些問題。圖1 4係假定可處理16色，32色，64色，128色模式之系統。於此場合，調色盤碼必須分別爲與色模式對應的4，5, 6，7位元》因爲：調色盤碼之長度=log2m (其中m =色模式數）例如在128色模式時，m = 128，故 l〇g2128 = log227 = 7 即需要7位元之寬度。因此，如把行進長度固定爲4位元，則如圖14所示壓縮資料之記錄長度變成可變長度。另方面，傳送資料之匯流排寬係依系統而定。雖然可變長資料本身並無問題，但於傳送資料時，必須配合電路之匯流排寬傳送。 -12- 本紙張尺度適用中國國家榇準（CNS ) A4規格（210X297公釐） (請先閱讀背面之注意事項再填寫本頁) T ，-=? 經濟部中央標牟局貝工消费合作社印製 A7 ___B7五、發明説明（1G ) 如把圖1 4之可變長資料’使用8位元匯流排寬之電路傳送時，如圖1 5所示必須依8位元單位送出資料。在 16色模式時，因傳送資料寬爲8位元，故不會發生問題，但如32色模式等非8之倍數的資料者，必須分割傳送。圖1 6表示把32色模式之圖像資料傳送時之情況。雖然圖像處理裝置產生之壓縮資料爲9位元，但由於在傳送給解碼裝置時必須分割成8位元送出，致使資料各偏移 1位元，使得一記錄跨到次一傳送資料。現在重複說明者，色模式係表示，可同時使用的有那些顏色。例如16色模式係表示可同時利用16種顏色。雖然色模式也依系統而異，但一般係如16(= 24)色模式，32(= 25)色模式，64(= 26)色模式，128 (= 27 )等以2的累乘決定彩色。此乃因表現彩色所必要的色資料尺寸爲此累乘値，例如1 6色模式時爲4位元，1 28色模式時爲7位元之故。如上述，色模式之顯示色表愈增，指定該色之色資料區域要愈長。然而依習知技術，1畫面只能指定1色模式。因之，即使畫面水平分割時顯示之色之數目依各領域而異之情況時，也必須以最多顯示色模式決定畫面模式。例如，欲作成由人工動畫及自然畫合成之畫面時，動畫只需少數的色（通常16色），但自然畫需要更多的色（通常16M色）。問題是，把這些圖像組合在一起時，整個畫面均需配合自然畫之色模式，無論以記憶體及資料傳送來說’都變 -13- 本紙張尺度適用中國國家梂準（CNS ) A4規格（2丨0X297公釐) (請先閲讀背面之注意事項再填寫本頁)1T Α7 Β7 V. Description of the Invention (5) The TV screen ends before displaying the 3rd and 16th raster. In this way, the rasters are processed in order to display one screen of data on the TV screen. This process is called normal regeneration. The image data expansion unit has a FIFO for image data from the control unit. When the FIFO is full, it sends a request signal of invalidation " DISABLE " (-REQR = 1) to the control order. When the control unit receives this signal, the data transmission is suspended. The control unit has an image data extension unit transmission control register as shown in FIG. 9, the image data extension unit data start address register and the transfer area block number register. The image data extension unit transmission control register is used to allow the transmitter of the image data extension unit data. If this permission is not obtained, no transmission will be made. If the transmission is not allowed during transmission, the transmission is stopped at this time. The start address register is used to designate the first address of the KRAM where the data like the data expansion unit is stored, and the control unit starts the transfer from the KRAM address. In block transfer, the address is automatically accumulated. # ί 浐部中心 ir.srxih-T 消于合竹 ίιyin (ti read the precautions on the back before filling in this Η) Transmission start The raster register is used to specify the control unit to start the image data expansion unit. The timing register for data transmission can specify the number of raster numbers. The transmission block number register is used to specify the number of transmission blocks when the 16 raster copies of data transmitted to the image data expansion unit are 1 block. If there is no change, each register will also display the same image in the next frame. Each register is valid after being set, but it is only valid when the block is being transferred to the image data expansion unit. Β Ψ nn * ϋ ^ — The actual transfer of data is limited to KB bus K- (BSU) was adjusted. ”This paper is compliant with the National Standard for CNS (CNS) A4 (210X297 mm). The Ministry of Justice " 4,-^ 力工消贤-? 竹右印 " A7 B7 V. Description of the Invention (6) Access to the image data expansion unit data is possible, and the image data expansion is performed with an output request signal (-REQR = 0). Data transmission becomes invalid in the following situations. (1) When the image data expansion unit has not started processing. (2) When the FIFO of the image data expansion unit is full. (3) Occasion of reading all the data of the number of bytes buried in the data during the period of counting HSYNC 16 times since the first data of the 16th line came. To stop the transmission of the control unit, the transmission start bit is disabled. As explained above, this system has 5 color modes of 16 colors, 32 colors, 64 colors, 128 colors, and 16M colors, and displays colors in YUV mode. Y represents brightness, and U and V represent color differences. Among them, 16 colors, 32 colors, 64 colors, and 128 colors are used as usual, and the 16M colors directly specify color information. The reason is to save memory size. For example, if you want to display 16 colors, you only need 4 bits, but if you use this 4 bits as color information directly, it will be fixed to 16 colors. Therefore, a color palette capable of specifying 512 copies of color data from 65,536 colors is prepared so that the displacement of the color palette can be specified for each frame, and the color palette within one frame can be selected with the above 4 bits. In this way, any of 16 colors can be selected and displayed from 65536 colors, so that flexibility can be produced. However, as mentioned above, in order to prepare a palette for 16M colors, a memory of several M bytes is required, which is difficult to be practical. Why do you need colored materials for 16M colors? 9- The size of this paper is described in the CNS Α4 size (2 丨 0X297 mm) 1 (诮 Please read the precautions on the back first, then this page) ,? τ Μ ^ '部屮次榀 ^-^ hT 消价合竹 ίι 卬 1'! A7 _________B7____ 5. Explanation of the invention (7) The reason for carrying the color information directly is here "The problem at this time is to indicate a point How many bits of color information you need to carry. For the largest number of 16M colors, it is necessary to have a size that can represent 16M colors. Therefore, to represent color information at one point, 24-bit β is required because 16M = 16 X IK xlK = 24 χ ΙΟ10 x 1〇10 = 224 is the reason, the 8-bit color data is assigned to Y, U, V to express color. The virtual picture in KRAM should be larger than the real picture. Because its size is 1024 x 1024 dots, if you want to prepare a virtual screen in 16M color mode, you need 24M bits of memory even if you use simple calculation. "But because the person using 16M color mode is usually a video recorder or image scanner The acquired natural portrait, so only the size of the picture is sufficient. In other words, 1 / 16th of the virtual picture is sufficient. Nonetheless, 5M bits are required. In the 16M color mode, in order to make the color data have an average of 16 bits per point, the table is as follows: (Y〇Y, U〇V〇, Y2 Y3 U, Vl5 ... Υη.2 Υη., Um In the VJ formula, m = (η-1) / 2 U and V of adjacent points are common, and the difference is represented by brightness Υ. As shown in Fig. 10, Υβ Y, uD V. can be divided into YQ U. ▽. And UQ V. There is still no problem with holding such image data, the reason is based on the fact that the color difference between adjacent points in natural portraits is not much. According to the above method, it can be converted to 32 bits. The color data indicates 2 points, and each point has an average of 16-bit color data, instead of 24-bit. 1 • 10- This paper is in accordance with the national standard (CNS) A4 specification (2 丨 OX297 (Chu) (诮 Read the precautions on the back before iAK-this page)-β A 7 __________ B7 V. Description of the invention (8) Only 1M bits are sufficient for the information on the screen. It can be seen that in the 16M color mode 1 vector factor corresponds to 2 points. This vector factor is sent from the control unit to the image data expansion unit in 16 raster units. In the case of a 1-to-1 vector factor, the read vector factor is The child can be sent in its original form, so it is easy to control. However, in the case of the vector factor of 1 to 2, this is not the case. Because the horizontally scrolling of the odd points will also send the non-displayed points. The vector factor is needed. The reason for reading the timing is here. To explain again, the video screen is composed of a collection of pixels (points). These pixels can be colored to create various images. At this time, the color of the color palette is used for each point. 》 In the processing of the computer, as shown in Figure 11, it is achieved by assigning the color code to the color data corresponding to the dot. The number of colors that can be used simultaneously in the color in the color palette is called Color mode. For example, when 16 colors are used at the same time, it is called 16-color mode. Generally, the TV screen used by computer game consoles is 256 X 240 dots. That is, one screen has 61,440 dots. In 16-color mode, the color information is The size requires 4 bits, so the size of the color data for 1 screen is: 61440 X 4 = 245760 bits = 30720 bytes = 30K bytes. The more the number of color modes, the larger the size of the color data. The amount of data has also become huge. It takes a lot of time to transmit this amount of data, which causes a big obstacle for video games that need to process images in a short time. Therefore, as shown in Figure 12 when data is transmitted, the data is compressed (coded), and the other is connected. -11 · The size of this paper is not compliant with the National Standards (CNS) Α4 size (HOX297) 诮 (Please read the precautions on the back first and then sacrifice this page) Order A7 B7 V. Description of Invention (9) Subject to this One party of the data decodes this data with a decoding device and displays the image. Since the color data of adjacent points of a general image are often the same, the image data is compressed in the form of < palette code > + < number >. Figure 13 shows the relationship between compressed data and points. When decoding, the number of color palette code (PLT) A is repeated to restore the original image data. This number is called the palette table (CRL). Therefore, this compression method is called the Run Length Method. According to the past, the color mode can only use 1 mode on 1 screen. Therefore, although < PLT > + < CRL > is a fixed length, no problem occurs. However, if multiple color modes can be specified for one screen, the fixed length < PLT > + < CRL > will cause some problems. Figure 14 is a system that assumes 16-color, 32-color, 64-color, and 128-color modes. In this case, the palette code must be 4, 5, 6, and 7 bits corresponding to the color mode. Because: The length of the palette code = log2m (where m = the number of color modes). For example, in the 128-color mode , M = 128, so 10g2128 = log227 = 7 requires a width of 7 bits. Therefore, if the travel length is fixed to 4 bits, the record length of the compressed data as shown in FIG. 14 becomes a variable length. On the other hand, the bus width for transmitting data depends on the system. Although the variable-length data itself is not a problem, when transmitting data, it must be transmitted with the bus width of the circuit. -12- This paper size applies to China National Standard (CNS) A4 (210X297 mm) (Please read the precautions on the back before filling out this page) T,-=? System A7 ___B7 V. Description of the invention (1G) If the variable-length data of Figure 14 is transmitted using an 8-bit bus width circuit, the data must be sent in 8-bit units as shown in Figure 15. In the 16-color mode, since the data width is 8 bits, no problem occurs. However, if the data is not a multiple of 8 such as the 32-color mode, it must be transmitted separately. Fig. 16 shows the situation when the image data in the 32-color mode is transmitted. Although the compressed data generated by the image processing device is 9 bits, it must be divided into 8 bits for transmission to the decoding device, which causes the data to be shifted by 1 bit, making one record cross the next transmission data. Now repeating the explanation, the color mode indicates those colors that can be used simultaneously. For example, the 16-color mode indicates that 16 colors can be used simultaneously. Although the color mode varies depending on the system, it is generally 16 (= 24) color mode, 32 (= 25) color mode, 64 (= 26) color mode, 128 (= 27) and so on. . This is because the size of the color data necessary for expressing color is multiplied by this. For example, it is 4 bits in 16-color mode and 7 bits in 28-color mode. As described above, the more the color table is displayed in the color mode, the longer the color data area for that color is specified. However, according to the conventional technology, only one color mode can be specified for one screen. Therefore, even if the number of colors displayed when the screen is divided horizontally varies from field to field, the screen mode must be determined in the maximum display color mode. For example, when you want to create a screen composed of artificial animation and natural painting, animation requires only a few colors (usually 16 colors), but natural painting requires more colors (usually 16M colors). The problem is that when these images are combined, the entire screen must match the color mode of natural painting, regardless of memory and data transfer. -13- This paper size applies to China National Standards (CNS) A4 Specifications (2 丨 0X297mm) (Please read the precautions on the back before filling in this page)

A7 B7 五、發明説明（n ) 成效率不良。鑑於上述，本發明係以開發在點與色向量非爲1對1 之情況時能實現水平捲動之方法爲目的。又，參考位元之位置改變，解碼裝置每於此改變時需要做資料之位置判定，以致增加其處理所耗之時間，換言之，增加硬體之負坦》因之，本發明之另一目的係要減輕解碼裝置之負擔。本發明之又一目的係要提供能以良好效率達成把色模式互不同之圖像表現於1畫面之圖像處理方法。解決問題之手段如上面之說明，由於來自控制單元之資料爲壓縮資料，故圖像資料擴展單元係把此資料回復原來的資料形式，並暫存於RRAM。然後從RRAM於每DCK週期讀出1向量單位，傳送給視頻編碼單元" 本發明係把此讀出時序及向視頻編碼單元之資料傳送時序控制，藉以施行水平捲動者。在內部準備尋取時序，以此時序讀取向量因子，即可把向量因子分爲亮度成分及色差成分，朝向視頻編碼單元傳送資料以行顯示。換言之，無水平捲動之向量因子讀取時序就是尋取時序。把此尋取時序偏移，即可實現水平捲動。不過，讀入時序也會依偶數點或奇數點而變。以下將進一步具體說明之。 -14- 本紙張尺度適用中國國家標準（CNS ) A4祝格（210X297公釐） (請先閲讀背面之注意事項再填寫本頁) 訂經濟部中央梯準局貝工消费合作社印装 Α7 Β7 經濟部中央標隼局貝工消費合作社印裝五、發明説明（12 ) 偶數時，以與水平捲動相同的時序讀取向量因子，並以同樣時序送出。此乃因讀取之資料全部是顯示資料之故。但讀取向量則不相同。例如，欲向畫面左方橫掃2點（水平捲動量+2)時， (Y2 Y3 Ui V, ' ——Υ„.2 Υ»-ι UB VB ' Ζ) 其中 m = (η-1)/2 爲讀取向量。從何資料讀起，可藉sex暫存器指定之。詳述之，在此暫存器設定相當於圖3所示BXR値（實際計數値）的數値，即可決定讀出之最先資料。實畫面範圍外之內容係依圓筒模式或非圓筒模式而異。圓筒模式時，內容爲YO Y1 uo V0，非圓筒模式即爲透明色資料。圓筒模式係謂，把橫掃出外之向量因子朝向畫之另一端繞回過來顯示之模式。奇數時，較之不做水平捲動時之讀取時序要偏移1時脈，俾使最先之向量因子不會碰到顯示期間（HDISP)。例如向畫面左方橫掃3點（水平捲動量+3)時， (Y2 Y3 U, V, ' Y4 Y5 U2 V2 > . . . . Υη.2 Um VB ' Ζ) 其中 m = (η+1 ) /2 爲傳送來之向量。此與向左捲動2點之情況相同，但藉讀入時序之位移，俾使最先之Y2 U, '不致侵犯到顯示期間，結果是施行了 3點的捲動。又，本發明係把色資料及行進長度做爲傳送資料長度 -15- (請先閲讀背面之注意事項再填寫本頁) -β 本紙張尺度適用中國國家標準（CNS ) Α4祝格（210Χ297公釐）經濟部中央標準局貝工消費合作社印製 A7 _B7_五、發明説明（13 ) 之整數倍長之固定長資料格式者。茲以圖17之圖畫爲例說明本發明。以16色模式顯示時，假設其顯示色呈如圖所示。單色模式時，則呈括弧內之黑白色圖畫。如從左方看圖中之直線a-a部份，於16色模式時，則呈藍（65)，紅（20)，白（15)，紅（40)，灰色（20)，綠（20)，灰色（60).:之顏色變化。但依單色模式時，則成爲白（65)，黑（20)，白（15)，黑（140)。括弧內之數値爲點數，亦即同時加以壓縮時之行進長度》由此可知，有顯示色愈多，行進長度則愈短之傾向。本發明係利用此現象，把資料之壓縮規則化如下。 (1) .壓縮長度爲固定長度，即等於資料匯流排寬或其整數 (2) .如顯示色模式之調色盤碼變長，即對稱地把行進長度縮短。 (3) .如行進長度之値過大致不能容納於固定長度時，則以固定長的行進長度爲零，並將行進長度指定於其次的記錄 (此稱擴充模式）。圖1 8係把上述之規則圖化者。若無（3)之規則，即使以同一色塡滿之圖像，亦會因行進長度短而必須作成許 #<PLT> + <CRL>資料，使壓縮效率低落。此外’爲要解決上述問題，本發明係於處理混含多數色模式之圖像資料之圖像處理裝置中，將色畫面分爲水平 -16- 本紙張尺度適用中^國家標率（CNS ) A4地格（210X297公釐） ί - —1- I - - I - - I I I --i - - -I ·1« In L i (請先閲讀背面之注意事項再填寫本頁) 經濟部中央標準局貝工消費合作社印製 A7 B7_五、發明説明（14 ) 遮沒單位，以各遮沒單位指定色模式，以傳送資料者。電視畫面之圖像顯示係如圖1 9所示，把掃瞄線橫掃於水平方向，使像素發色。因之，本發明係把畫面分爲幾個等分，並按分割之每個畫面決定色模式。現在假定以η 光柵單位分割畫面。此時，把η光柵份稱爲ηΗ(水平遮沒）。本發明係以下列方式決定每ηΗ之色模式。 (1 )檢測ηΗ所含之色模式種類。 (2) 如有多種色模式時，則以其中色數最多之色模式爲其區塊之色模式。 (3) 配合色模式改寫ηΗ內之色資料。依上述方式，可決定每ηΗ之色模式及每像素之色碼 (顯示色）。當此資料向他裝置傳送時，如圖2 0所示，在傳送資料之開頭附上標誌碼，其後方連續以每像素之色資料。實施例：依本發明之系統，畫面顯示控制信號除了 VSYNC(V遮沒），HSYNC(H遮沒）之中斷信號外，亦有DCK。 HSYNC信號係從視頻編碼單元發送至各裝置。中斷處理乃根據此信號而起動，控制單元即把以16光柵單位壓縮之圖像資料送至圖像資料擴展單元。圖像資料擴展單元接收此資料並回復爲原資料形式，暫存於RRAM。 DCK爲1點單位之中斷信號，藉此信號始能每於顯示 1個點，則施行中斷1次，以便施行點單位之操作。本發 •17- 冢ϋ尺度適用中國國家標覃（CNS ) A4祝格U10X297公釐） - (請先閲讀背面之注_項再填寫本頁) 訂經濟部中央梯準局員工消費合作社印裝 A7 B7五、發明説明（15 ) 明也利用此DCK。嗣後，把DCK之時鐘週期簡稱時鐘。現在舉出無水平捲動，向左捲動1個點，及向右捲動 1個點的3個例子。由於捲動幾個點係依讀取之最初向量因子（參閱前述之2點，3點之例子）而定，故以後之偶數點捲動或奇數點捲動，均可依現在要舉出之時序而解決之。圖2 1，圖2 2及圖2 3分別爲水平捲動量爲零（無捲動），+1(左掃1點），-1(右掃1點）之讀取時序之例子。水平捲動量爲零時，如圖2 1所示配合尋取時序讀取 Y〇 Y, U。V。，分爲YQ, Y,，U。，V。，送給配合顯示期間之視頻碥碼單元。以最初點（像素）被顯示之時序，藉緩衝器讀取其次之Y2 Y3比V,。重複此操作即可施行通常的圖像顯示。水平捲動量爲+ 1時，如圖2 2所示，在尋取時序之2 時鐘前讀取Y。Y, U。V。。藉此，可將最初的YQ u〇 Vo納入緩衝器，但由於不在顯示期間內，致變成無效信號，不會被顯示。因此，最初被顯示的係自Y, UQ VD起。亦即，往左偏移1個點。於此場合，最後的點會成爲無效（invalid)，但在圓筒模式時’把最初之Υ(5 W ν〇傳送至視頻編碼單元。非圓筒模式時’即送出無效信號。換言之，非圓筒模式時最後的點成爲透明色。水平捲動量爲-1時，如圖2 3所示，較尋取時序要慢2時鐘讀取。藉此，第1個點者變成無顯示， • 18- 本ϋ尺度適用中囷國ϋ準（) Μ样^ ( 210X297公釐) (請先閲讀背面之注意事項再填寫本頁) 訂經濟部中夬樣準局貝工消費合作社印製 A7 ___B7___五、發明説明（16 ) 第2個起顯示YO UO V0。換言之，向右偏移1個點。不過，上述之例子係非圓筒模式之情況。於圓筒模式時，以與水平捲動量爲+1同樣的時序讀取最後之向量因子Yn.2 Yu Um V„，即可把Yu un v„顯示於最初之第1個點。也就是說，自第2點起，從Yo Y, u0 V〇之通常順序讀取就可。圖8〜1 0係以讀取之時序爲中心而看者，圖2 4~2 6係以向視頻編碼單元送出之時序爲中心而看的圖。向視頻編碼單元傳送之資料係以YYUV單位，向YUV 之解碼係以視頻編碼單元施行。儘管以YYUV傳送，也不致顯示不在顯示期間之YUV。本發明係利用此時序來實現捲動。圖 2 4 ~2 6 中，RT0-RT7 表示匯流排，RTC0DE1-RTCODEO表示匯流排上之圖像資料之種類。圖2 7表示圖像資料擴展單元與視頻編碼單元之介面》圖28爲上述水平捲動例子之綜合。圖中，m爲（11-1)/2，INVALID爲非顯示（透明色顯示），「讀出時序」表示最初讀取之向量因子對於尋取時序之快慢度，分別以「+」稍決，「-」稍慢進行讀取。記憶體上之YYUV之配置示於圖2 9。依本發明，藉圖像資訊讀出時序及送出時序之控制，可平滑施行水平捲動。藉此，對於構成圖像之圖像發色所用之向量因子資訊與點未必然爲1對1之情況，亦能平滑施行捲動。 -19- 本紙張尺度適用中國國家標準（CNS ) Α4ϋ格（210X297公釐） (請先閲讀背面之注意事項再填寫本頁) 經濟部中央標準局貝工消费合作社印装 A7 B7五、發明説明（17 ) 因此，可使資料賦有把如自然圖像般豐富色資訊之色資料壓縮之形式，從而可增進記億體之節省，傳送速度之高速化等之功效。 . 接之，說明本發明第2實施例，圖像處理裝置。圖6 爲其方塊圖。如前所述，其構成爲以CD-ROM等之遊戲軟體記錄媒體，32位元CPU，圖像聲音傳送控制及各裝置之介面爲主的控制單元，圖像資料擴展單元，圖像資料輸出單元，聲音資料輸出單元，視頻編碼單元、VDP單元等。亦有各單元專用之如K-RAM, M-RAM, P-RAM, V-RAM之記億體。， CPU具有透過記億體支援直接控制DRAM之記億體控制功能及透過I/O埠與種種周邊機器通信之I/O控制功能。亦有定時器，平行輸入出埠及中斷控制功能。 VDP把CPU所寫入於VRAM之顯示資料讀出，送至視頻編碼單元，以在畫面顯示出來》VDP單元可獲得最大2 個畫面，即如以往由8 X 8區塊之外部區塊序列形式之背景畫面及子畫面（SPRITE)組合的畫面。圖3 0表示控制單元之方塊圖2控制單元內藏有SCSI 控制器，係從如CD-RAM等外部記億裝置經由SCSI介面取入圖像或聲音等之資料。所取入之資料先暫在K-RAM緩衝。控制單元內藏有DRAM控制器，用以以預定之時序讀出K-RAM所存之資料》自然畫背景圖像資料係在控制單元內受到1點資料單元之優先級判定後送至視頻編碼單元。 -20· 本紙張尺度適用中國國家揉準（CNS > A4規格（210X297公釐） " (請先閲讀背面之注意事項再填寫本頁) 經濟部中央橾準局貝工消費合作杜印裝 A7 B7__五、發明説明（18 ) 已被資料壓縮之動畫（全彩，調色盤）資料被送至圖像資料擴展單元。圖像資料擴展單元爲如前述之圖7之方塊圖所示。此擴展單元把資料擴展後送至視頻編碼單元。圖像資料擴展單元可處理自然畫動畫用之CDT轉換，哈夫曼碼化方式壓縮資料，及人造動畫用之行進長度碼化方式壓縮資料。亦備有反DCT轉換手段*反量子化手段，哈夫曼編碼解碼手段及行進長度編碼解碼手段。視頻編碼單元把來自VDP單元，控制單元，圖像資料擴展單元之VDP圖像，自然畫背景圖像，動畫（全彩，調色盤）資料加以疊合處理，施加調色盤再生，特殊效果處理水及D/A轉換等之處理後予以輸出，再藉外部電路輸出最後編碼成NTSC信號之圖像信號。從CD-ROM等讀入之ADPCM聲音資料係與圖像資料一樣在KRAM緩衝後，以控制單元送至聲音資料輸出單元以便再生。依此實施例之裝置，調色盤彩色可處理16色，32色， 64色，128色模式。圖像資料之傳輸係以16光栅（線）單位進行，匯流排寬爲8位元。圖3 1表示其資料製作方式。在本系統中雖可使多個色模式混合存在於畫面，但在圖3 1所示之16光柵單位，其畫面爲同一色模式》此外，A爲圖像資料之類型。如A有指定FFH，F8H 者，就是IDCT用壓縮資料。這是處理自然畫時之圖像資料。A爲F3H，F2H，F1H，F0H者，是附調色盤之圖像資料之情況，分別表示128色，64色，32色，16色行進長 -2卜本紙張尺度適用中國國家標準（CNS ) 格（2丨0X297公釐） " (請先閲讀背面之注$項再填寫本页) 訂 A 7 _B7_ 五、發明説明（19 ) 「-•%先開讀背面之注意事項再填寫本頁〕度用壓縮資料，可利用於人造動畫UxH爲16進位之XX) β Β表示壓縮資料領域之資料長上位位元組，C表示下位位元組，D表示把壓縮資料予以2位元組邊界處理所用之資料領域。壓縮資料領域指定有調色盤碼及行進長度。其格式示於圖3 2。是否使用擴充模式·，由用戶程式加以判定。例如，在64色模式時，因行進長度（CRL)爲2位元，故最高只能示出4之値。因之，如調色盤碼13之f料連續100個時，則作爲擴充模式，並指定如下：最初之資料：<13> + <0>......長度8位元其次之資料：<100> ........長度8位元解碼裝置係於調色盤碼之接下來之値爲零時判斷有其次記錄之行進長度，並以100作爲行進長度。即從解碼裝置之輸出，生成100個調色盤碼13之資料，送至視頻編碼單元，以行圖像顯示。現在看擴充模式之具體例。64色模式，調色盤碼13, 行進長度10之場合，若不使用擴充模式時，則需要 <13> + <4> 經濟部中央橾準局貝工消費合作社印製之資料25個。因<13> + <4>之長度爲8位元，總共需要200位元》反之*若利用擴充模式時，僅需 <13> + <0> <100> 之2個資料就可。因都是8位元，總共16位元即夠。擴 -22- 本紙張尺度適用中國國家揉準（CNS ) A4祝格（210X297公釐）經濟部中央標準局負工消费合作社印装 A7 ΒΊ_五、發明説明（20 ) 充模式之重要性在此。依本發明，由於壓縮資料不是等於匯流排寬就是其整數倍，可輕易準備送給解碼裝置之資料。從而能輕易編寫程式。也可簡化解碼裝置之解碼邏輯，減輕解碼裝置之負擔，提高解碼速度等效果。茲將本發明第3實施例之圖像處理裝置之槪略說明如下。整個裝置之方塊圖示於第6圖。其係由以CD-ROM等遊戲軟體記錄媒體，32位元CPU，圖像聲音資料傳送控制與各裝置之介面爲主之控制單元，圖像資料擴展轉換單元，圖像資料輸出單元，聲音資料輸出單元，視頻編碼單元及VDP單元等所構成。爲了供各單元之專用，具有K-RAM, Μ·RAM, V-RAM等之記憶體。 CPU具有透過記憶體支援可直接控制DRAM之記憶體控制功能及透過I/O埠而與種種周邊機器通信之I/O控制功能》也具有定時器，平行輸入出埠及中斷控制功能。 VDP單元把CPU所寫入於VRAM之顯示資料，送至視頻編碼單元，藉以顯示在畫面。VDP單元可獲得最大2個如以往8 X 8區塊之外部區塊序列形式之由背景畫面及子畫面組成之畫面。控制單元之方塊圖示於前述之圖3 0。控制單元內藏有SCSI控制器，以便經由SCSI介面從CD-ROM等之外部記憶裝置取入圖像，聲音等之資料。所取入之資料暫緩衝存入K-RAM。控制單元內藏有DRAM控制器，利用其作用把存於K-RAM之資料依預定之時序讀出。 -23- 本紙張尺度通用中國國家標準（CNS ) A4規格（2!0><297公釐) (請先閲讀背面之注意事項再填寫本頁)A7 B7 5. Description of the invention (n) Poor production efficiency. In view of the foregoing, the present invention aims to develop a method capable of achieving horizontal scrolling when the point and color vectors are not one to one. In addition, the position of the reference bit changes, and the decoding device needs to make a position determination of the data each time it changes, so as to increase the time it takes to process it, in other words, to increase the negative load of the hardware. Therefore, another object of the present invention This is to reduce the burden on the decoding device. Still another object of the present invention is to provide an image processing method capable of achieving, with good efficiency, displaying images with mutually different color modes on one screen. Means to solve the problem As explained above, since the data from the control unit is compressed data, the image data expansion unit restores this data to the original data form and temporarily stores it in RRAM. Then read 1 vector unit from the RRAM every DCK cycle and send it to the video coding unit. The present invention is to control the timing of this readout and the timing of data transmission to the video coding unit to implement horizontal scrolling. Internally prepare to find the timing, and read the vector factor at this timing. The vector factor can be divided into the luminance component and the color difference component, and the data is transmitted to the video encoding unit for line display. In other words, the read timing of the vector factor without horizontal scrolling is the seek timing. By shifting this seek timing, horizontal scrolling can be achieved. However, the read-in timing also varies depending on the even or odd points. This will be explained in more detail below. -14- This paper size applies the Chinese National Standard (CNS) A4 Zhuge (210X297mm) (Please read the notes on the back before filling this page) Order printed by the Central Laboratories of the Ministry of Economic Affairs and printed by the Bayer Consumer Cooperative Α7 Β7 Economy Printed by the Ministry of Standards and Technology Bureau, Shellfish Consumer Cooperatives. 5. Description of the Invention (12) When the number is even, the vector factor is read at the same timing as the horizontal scrolling and sent at the same timing. This is because the read data is all display data. But read vectors are different. For example, if you want to swipe 2 points to the left of the screen (horizontal scrolling amount +2), (Y2 Y3 Ui V, '——Υ „.2 Υ» -ι UB VB' Z) where m = (η-1) / 2 is the reading vector. From which data can be read, it can be specified by the sex register. In detail, the register is set to a number equal to BXR 値 (actual count 値) shown in Figure 3, and Determine the first data to be read. The content outside the real screen range depends on the cylindrical mode or non-cylindrical mode. In the cylindrical mode, the content is YO Y1 uo V0, and the non-cylindrical mode is transparent data. Circle The tube mode refers to a mode in which the vector factor swept out is turned toward the other end of the painting and displayed. When the number is odd, the reading timing is shifted by 1 clock than when the horizontal scrolling is not performed, making the first The vector factor does not touch the display period (HDISP). For example, when you swipe 3 points (horizontal scroll amount +3) to the left of the screen, (Y2 Y3 U, V, 'Y4 Y5 U2 V2 >... Υη.2 Um VB 'ZZ) where m = (η + 1) / 2 is the vector transmitted. This is the same as the case of scrolling 2 points to the left, but by reading the timing shift, the first Y2 U,' Inviolable During the display period, the result was scrolling at 3 points. In addition, the present invention uses color data and travel length as transmission data length -15- (Please read the precautions on the back before filling this page) -β This paper The standard applies to the Chinese National Standard (CNS) Α4 Zhuge (210 × 297 mm) The printed format of the fixed-length data of integer multiples of the invention description (13) printed by the A7 _B7_ of the Beige Consumer Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs. The drawing in Fig. 17 is taken as an example to illustrate the present invention. When displaying in the 16-color mode, it is assumed that the display color is as shown in the figure. In the monochrome mode, it is a black and white drawing in brackets. Part aa, in the 16-color mode, it is blue (65), red (20), white (15), red (40), gray (20), green (20), and gray (60) .: However, in the monochrome mode, it becomes white (65), black (20), white (15), and black (140). The number in parentheses is the number of points, which is the travel length when compressed at the same time. " It can be seen from this that the more the display color, the shorter the travel length. The present invention uses this phenomenon to make The compression rules of the materials are as follows: (1). The compression length is a fixed length, which is equal to the data bus width or its integer (2). If the color palette code of the display color mode becomes longer, the travel length is shortened symmetrically. (3). If the length of the travel length cannot be accommodated in a fixed length, the fixed length travel length is zero, and the travel length is designated as the next record (this is called the extended mode). Figure 18 shows the above. If there is no rule (3), even if the image is filled with the same color, it will have to be created because of the short travel length. # &Lt; PLT > + < CRL > . In addition, in order to solve the above-mentioned problems, the present invention is an image processing device that processes image data containing a plurality of color modes, and divides the color screen into horizontal -16- This paper scale is applicable ^ National standard rate (CNS) A4 grid (210X297 mm) ί-—1- I--I--III --i---I · 1 «In L i (Please read the precautions on the back before filling this page) Central Standard of the Ministry of Economic Affairs A7 B7 printed by Bureau Shelley Consumer Cooperative Co., Ltd. V. Description of the invention (14) Obscured units, designated color mode for each obscured unit, to transmit data. The image display of the TV screen is shown in Figure 19, scanning the scanning line horizontally to make the pixels color. Therefore, the present invention divides the picture into several equal parts, and determines the color mode for each divided picture. It is now assumed that the picture is divided in n raster units. At this time, the η grating component is called ηΗ (horizontal occlusion). The present invention determines the color mode per ηΗ in the following manner. (1) Detect the type of color mode contained in ηΗ. (2) If there are multiple color modes, the color mode with the largest number of colors is the color mode of its block. (3) Rewrite color data in ηΗ with color mode. In the above manner, the color mode per ηΗ and the color code (display color) per pixel can be determined. When this data is transmitted to other devices, as shown in FIG. 20, a logo code is attached to the beginning of the transmission data, and the color data of each pixel is continuous behind it. Embodiment: According to the system of the present invention, in addition to the interrupt signals of VSYNC (V mask) and HSYNC (H mask), the picture display control signal also has DCK. The HSYNC signal is sent from the video encoding unit to each device. The interrupt processing is started based on this signal, and the control unit sends the image data compressed in 16 raster units to the image data expansion unit. Image data extension unit receives this data and returns it to the original data form, temporarily stored in RRAM. DCK is an interrupt signal in units of 1 point. With this signal, every time a point is displayed, an interrupt is performed once to perform operations in points. This issue • 17- The standard of gravel ϋ is applicable to Chinese National Standard Tan (CNS) A4 Zhuge U10X297 mm)-(Please read the note on the back before filling this page) Order printed by the Central Consumers ’Cooperative Bureau of the Ministry of Economic Affairs A7 B7 V. Description of Invention (15) The DCK is also used. After that, the clock cycle of DCK is referred to as clock. Here are three examples of scrolling horizontally, scrolling one point to the left, and scrolling one point to the right. Since the scrolling of several points depends on the initial vector factor read (see the example of 2 points and 3 points above), the scrolling of even-numbered points or odd-numbered points in the future can be listed as follows. And timing. Figure 21, Figure 2 2 and Figure 2 3 are examples of reading timings when the horizontal scrolling amount is zero (no scrolling), +1 (left sweep 1 point), -1 (right sweep 1 point). When the horizontal scrolling amount is zero, as shown in FIG. 21, read Y0, Y in conjunction with the seek timing. V. , Divided into YQ, Y ,, U. , V. To the video code unit during the display period. At the timing when the first point (pixel) is displayed, read the second Y2 Y3 ratio V, by the buffer. Repeat this operation for the usual image display. When the horizontal scrolling amount is +1, as shown in Figure 22, Y is read before the second clock of the timing is found. Y, U. V. . In this way, the original YQ u〇 Vo can be incorporated into the buffer, but because it is not in the display period, it becomes an invalid signal and will not be displayed. Therefore, the first displayed lines are from Y, UQ VD. That is, it is shifted to the left by 1 point. In this case, the last point will be invalid, but in the cylindrical mode, 'the first Υ (5 W ν0) is transmitted to the video encoding unit. In the non-cylindrical mode, an invalid signal is sent. In other words, the non- In the cylindrical mode, the last point becomes transparent. When the horizontal scrolling amount is -1, as shown in Figure 23, it is 2 clocks slower than the search timing. With this, the first point becomes no display, • 18- This standard is applicable to the Chinese Standard (M) sample ^ (210X297 mm) (Please read the precautions on the back before filling out this page) Order printed by the China Standard Sample Bureau of the Ministry of Economic Affairs Printed by Shellfish Consumer Cooperatives _B7__ _Fifth, the description of the invention (16) YO UO V0 is displayed from the second one. In other words, it is shifted to the right by 1 point. However, the above example is the case of the non-cylindrical mode. Read the last vector factor Yn.2 Yu Um V „at the same timing with momentum +1, and you can display Yu un v„ at the first point. That is, starting from point 2, starting from Yo Y , u0 V〇 can be read in the usual order. Figures 8 ~ 10 are centered on the timing of reading, and Figures 2 ~ 2-6 are based on The picture sent to the video encoding unit is centered on the timing. The data transmitted to the video encoding unit is implemented in YYUV units, and the decoding of YUV is performed in the video encoding unit. Although transmitted in YYUV, it does not display the YUV that is not in the display period. The present invention uses this sequence to implement scrolling. In Figures 2 to 26, RT0-RT7 represents the bus, and RTC0DE1-RTCODEO represents the type of image data on the bus. Figure 2 7 represents the image data expansion unit Interface with Video Coding Unit "Figure 28 is a synthesis of the above horizontal scrolling example. In the figure, m is (11-1) / 2, INVALID is non-display (transparent color display), and" read timing "indicates the first read The speed of the vector factor for the search timing is determined by "+", and "-" is slightly slower to read. The YYUV configuration on the memory is shown in Figure 29. According to the present invention, the image information is read The control of the output timing and the output timing can smoothly perform horizontal scrolling. Thus, for situations where the vector factor information and points used for the color development of the image constituting the image are not necessarily 1 to 1, scrolling can also be performed smoothly. -19- this paper Applicable to China National Standard (CNS) Α4ϋ 格 (210X297 mm) (Please read the notes on the back before filling this page) Printed by the Central Standards Bureau of the Ministry of Economic Affairs, Shellfish Consumer Cooperatives, printed A7 B7 V. Invention Description (17) Therefore, The data can be given a form of compressing color data that is rich in color information like a natural image, which can increase the effects of saving billions of records and speeding up the transmission speed. Next, the second embodiment of the present invention will be described. Image processing device. Figure 6 is a block diagram. As mentioned earlier, it is composed of a game software recording medium such as a CD-ROM, a 32-bit CPU, image and sound transmission control, and control based on the interface of each device. Unit, image data extension unit, image data output unit, sound data output unit, video encoding unit, VDP unit, etc. There are also hundreds of millions dedicated to each unit such as K-RAM, M-RAM, P-RAM, V-RAM. The CPU has a memory controller function that directly controls the DRAM through the memory controller and an I / O control function that communicates with various peripheral devices through the I / O port. There are also timers, parallel input and output ports and interrupt control functions. The VDP reads the display data written by the CPU in the VRAM and sends it to the video encoding unit to display it on the screen. The VDP unit can obtain a maximum of 2 pictures, which is the same as the external block sequence of 8 X 8 blocks. Background picture and sprite (SPRITE) combination picture. Fig. 30 shows the block of the control unit. Fig. 2 The control unit contains a SCSI controller, which is used to retrieve data such as images or sounds from an external device such as CD-RAM through a SCSI interface. The acquired data is temporarily buffered in K-RAM. The control unit contains a DRAM controller to read the data stored in the K-RAM at a predetermined timing. Natural background image data is sent to the video encoding unit after being judged by the priority of the 1-point data unit in the control unit. . -20 · The size of this paper is applicable to the Chinese national standard (CNS > A4 size (210X297mm) " (Please read the precautions on the back before filling this page). A7 B7__V. Description of the invention (18) The compressed data (full color, color palette) data has been sent to the image data expansion unit. The image data expansion unit is as shown in the block diagram of FIG. 7 described above. This expansion unit expands the data and sends it to the video encoding unit. The image data expansion unit can handle the CDT conversion of natural drawing animation, Huffman coding method to compress the data, and the artificial animation using the length length coding method to compress the data .Also equipped with anti-DCT conversion method * anti-quantization method, Huffman encoding and decoding method, and travel length encoding and decoding method. The video encoding unit converts VDP images from the VDP unit, control unit, and image data expansion unit into natural pictures. The background image, animation (full color, color palette) data are superimposed, and the color palette regeneration, special effects processing water and D / A conversion are processed and output, and then borrowed. The circuit outputs the image signal finally encoded into NTSC signal. The ADPCM sound data read from CD-ROM etc. is buffered in KRAM like the image data, and then sent to the sound data output unit by the control unit for reproduction. According to this embodiment With this device, the color palette can process 16-color, 32-color, 64-color, and 128-color modes. The transmission of image data is performed in 16 raster (line) units, and the bus width is 8 bits. Figure 3 1 shows its Data production method. Although multiple color modes can be mixed in the screen in this system, the 16 raster units shown in Figure 31 are the same color mode. In addition, A is the type of image data. A has the specified FFH, F8H, which is the compressed data for IDCT. This is the image data when processing natural paintings. A is F3H, F2H, F1H, F0H, which is the case of image data with palette, respectively. 128-color, 64-color, 32-color, 16-color travel length-2 paper size Applicable to the Chinese National Standard (CNS) grid (2 丨 0X297 mm) " (Please read the note on the back before filling this page) Order A 7 _B7_ V. Description of the invention (19) "-•% read first Please fill in this page again before filling in this page.] The compressed data can be used for artificial animation UxH is hexadecimal XX) β Β means the data in the compressed data field is the upper byte, C is the lower byte, D is the Compressed data is a data field used for 2-byte boundary processing. The compressed data field is specified with a palette code and a travel length. The format is shown in Figure 3 2. Whether to use expansion mode or not is determined by the user program. For example, in In 64-color mode, because the travel length (CRL) is 2 digits, only the maximum of 4 can be shown. Therefore, if there are 100 consecutive f materials in the palette code 13, it is used as the expansion mode and specified As follows: Initial data: < 13 > + < 0 > ...... 8-bit length. Second data: < 100 > ........ The 8-bit length decoding device is used for tuning. When the next one of the disc code is zero, it is judged that there is the next recorded travel length, and 100 is used as the travel length. That is, from the output of the decoding device, 100 pieces of data of the color code 13 are generated and sent to the video encoding unit for display as a line image. Now look at a specific example of the expansion mode. For 64-color mode, color palette code 13, and travel length 10, if you do not use the extended mode, you need <13> + + 4> 4 materials printed by the Shellfish Consumer Cooperative of the Central Bureau of Standards, Ministry of Economic Affairs . Because the length of < 13 > + < 4 > is 8 bits, a total of 200 bits is required. Conversely * when using the extended mode, only 2 data of < 13 > + < 0 > < 100 > Just fine. Since they are all 8 bits, a total of 16 bits is enough. Extend-22- This paper size applies to China National Standards (CNS) A4 Zhuge (210X297 mm) Printed A7 by the Central Standards Bureau of the Ministry of Economic Affairs and Consumer Cooperatives Ⅴ. Explanation of the invention (20) The importance of the charging mode is this. According to the present invention, since the compressed data is either equal to the bus width or an integer multiple thereof, it is easy to prepare the data to be sent to the decoding device. This makes it easy to write programs. It can also simplify the decoding logic of the decoding device, reduce the burden on the decoding device, and improve the decoding speed. The outline of the image processing apparatus according to the third embodiment of the present invention will be described below. The block diagram of the entire device is shown in Figure 6. It is a control unit mainly based on game software recording media such as CD-ROM, 32-bit CPU, image and sound data transmission control and the interface of each device, image data extension conversion unit, image data output unit, and sound data. It consists of output unit, video encoding unit and VDP unit. For the exclusive use of each unit, it has K-RAM, M · RAM, V-RAM and other memories. The CPU has a memory control function that can directly control the DRAM through the memory and an I / O control function that communicates with various peripheral devices through the I / O port. It also has a timer, parallel input and output ports, and interrupt control functions. The VDP unit sends the display data written in the VRAM by the CPU to the video encoding unit for display on the screen. The VDP unit can obtain a maximum of 2 pictures consisting of a background picture and a sub picture in the form of an external block sequence of the previous 8 X 8 block. The block diagram of the control unit is shown in Figure 30 above. The control unit has a built-in SCSI controller to retrieve image, sound, and other data from external storage devices such as CD-ROMs via the SCSI interface. The acquired data is temporarily buffered and stored in K-RAM. The control unit contains a DRAM controller, which uses its function to read the data stored in the K-RAM at a predetermined timing. -23- The paper size is in accordance with Chinese National Standard (CNS) A4 specification (2! 0 > < 297 mm)

、1T A7 B7 經濟部智慧財產局員工消費合作社印製五、發明説明（21 ) 自然畫背景圖像資料在控制單元內以1點資料單位接受優先級判定後送至視頻編碼單元。已資料壓縮之動畫像（全彩，調色盤）資料被送至圖像資料擴展單元。前述之圖7示出此擴展單元之方塊圖。此擴展單元把資料擴展後送至視頻編碼單元。圖像資料擴展單元可處理自然畫動畫用之DCT轉換及哈夫曼編碼方式壓縮資料，人造動畫用之行進長度編碼方式壓縮資料。又備有反DCT轉換手段，反量子化手段，哈夫曼解碼手段，行進長度解碼手段。視頻編碼單元係對來自VDP單元，控制單元*圖像資料擴展單元之VDP圖像，自然畫背景圖像，動畫像（全彩，調色盤）資料施加疊合處理，調色盤再生*特殊效果處理及D/A轉換等之處理並予以輸出，再藉由外部電路把最後編碼成NTSC信號之圖像信號輸出。 CD-ROM等所讀入之ADPCM聲音資料係與圖像資料同樣先在KRAM緩衝後，由控制單元送至聲音資料輸出單元，以便再生。茲就圖像之再生詳加說明。控制單元從CD-ROM讀取圖像資料，暫存於K-RAM。再依需要把資料從RAM取出，經旋轉，放大等之圖像處理後傳送至下一個處理階段。茲就不同色模式混合存在之圖像的顯示說明如下。本發明係把畫面分爲16H單位，附加標誌碼，再把已壓縮之色資料傳送至圖像資料擴展單元。擴展單元把壓縮色資料利用解碼裝置回復原形，並配合顯示時序傳送至視頻編碼 -24- 本紙張尺度適用中國國家標隼（CNS > A4規格（210X297公釐) (請先閲讀背面之注意Ϋ項再填寫本頁) 訂, 1T A7 B7 Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. 5. Description of the invention (21) The background image data of natural painting is sent to the video coding unit after receiving priority determination in the control unit with 1 point data unit. The data of the compressed image (full color, color palette) is sent to the image data expansion unit. The aforementioned FIG. 7 shows a block diagram of the expansion unit. This expansion unit sends the data to the video coding unit after expansion. The image data expansion unit can handle DCT conversion for natural drawing animation and Huffman coding compression data, and marching length coding for artificial animation compression data. It also has anti-DCT conversion means, dequantization means, Huffman decoding means, and travel length decoding means. The video encoding unit applies superposition processing to the VDP image, natural drawing background image, and animation image (full color, color palette) data from the VDP unit, control unit * image data expansion unit, and palette reproduction * special The effect processing and D / A conversion are processed and output, and then the image signal finally encoded into NTSC signal is output by external circuit. The ADPCM audio data read by CD-ROM and the like is first buffered in KRAM and sent by the control unit to the audio data output unit for reproduction, as is the image data. The reproduction of the image is explained in detail. The control unit reads the image data from the CD-ROM and temporarily stores it in K-RAM. Then take the data out of the RAM as needed, and then transfer it to the next processing stage after image processing such as rotation and enlargement. The following describes the display of images in which different color modes are mixed. The invention divides the picture into 16H units, attaches a mark code, and transmits the compressed color data to the image data expansion unit. The expansion unit restores the compressed color data to its original form using the decoding device, and transmits it to the video code with the display timing. -24- This paper size applies to the Chinese national standard (CNS > A4 size (210X297 mm) (Please read the note on the back first) (Please fill this page again)

I..V A7 B7 五、發明説明（22 ) 器。此時，色模式係視附加在16H單位之頭標之標誌碼而變換。如無標誌碼，則不變換色模式。換句話說，繼續進行既定之模式。視頻編碼器把其色資料傳送至電視裝置以行圖像顯示。本發明之圖像處理裝置，除了人造動畫像外，亦可處理錄影機或影像掃瞄機取入之自然圖像。這些圖像可顯示於各別的BG畫面（背景畫面本系統可處理各自獨立的 4個BG畫面，控制單元係於向圖像資料擴展單元送出以前，把這些獨立BG畫面合成。以自然圖像而言，大約可顯示1600萬色（16M色模式）。顯示16M色所需之位元數，因爲16M爲： 16M = 16 X IK X IK = 24 X ΙΟ10 χ ΙΟ10 = 224 所以，是24位元（=3位元組）一般電視畫面爲256 χ 240點（像素）之構成。現在假定如圖3 3所示，畫面被地平線2分割爲天空及地面之情況。畫面上部爲人造動畫像之藍色天空及綠色山岳，畫面下部爲自然圖像，有穿著豐富彩色衣服之人物等存在，如圖3 4所示，此畫像可藉BG畫面合成製作之。如使用單一色模式，則必須以1 6M色模式處理整個畫面，故顯示1個畫面份色資料所需之容量爲： 256 χ 240 χ 3 = 184，320 = 180K 位元組因之，如圖3 3所示，假定在畫面上部使用16色模式，畫面下部使用16M色模式。再假定色模式類型有8種， -25- 本紙張尺度適用中國國家標準（CNS ) A4規格（2丨0><297公釐） (請先閱讀背面之注意事項再填寫本頁)I..V A7 B7 5. Invention Description (22). At this time, the color mode is changed depending on the flag code attached to the header of the 16H unit. If there is no logo code, the color mode is not changed. In other words, continue with the established model. The video encoder transmits its color data to a TV device for image display. In addition to the artificial animation image, the image processing device of the present invention can process natural images taken by a video camera or an image scanner. These images can be displayed on separate BG screens (background screens. This system can process 4 independent BG screens. The control unit is to synthesize these independent BG screens before sending them to the image data expansion unit. Natural images For example, it can display about 16 million colors (16M color mode). The number of bits required to display 16M colors, because 16M is: 16M = 16 X IK X IK = 24 X ΙΟ10 χ ΙΟ10 = 224 So, it is 24 bits (= 3 bytes) The general television picture is composed of 256 x 240 dots (pixels). Now suppose that the picture is divided by the horizon 2 into the sky and the ground as shown in Fig. 33. The upper part of the picture is the blue of the artificial animation image. Color sky and green mountains, the lower part of the screen is a natural image, there are people wearing rich colored clothes, as shown in Figure 34, this portrait can be synthesized by BG screen. If using a single color mode, you must use 1 The 6M color mode processes the entire screen, so the capacity required to display the color data of one screen is: 256 χ 240 χ 3 = 184, 320 = 180K bytes. Therefore, as shown in Figure 3, it is assumed that it is used in the upper part of the screen. 16 color mode, draw The lower part uses 16M color mode. It is also assumed that there are 8 types of color modes. -25- This paper size applies to China National Standard (CNS) A4 specifications (2 丨 0 > < 297 mm) (Please read the precautions on the back first) (Fill in this page)

.OT 經濟部智慧財產局員工消費合作社印製經濟部智慧財產局員工消費合作社印製 A7 __B7_五、發明説明（23 ) 並按每1個像素單位施行模式變換之情況。因爲表示色模式之標誌碼需要3位元，故表示1個像素所必要之位元數即爲： 16M色模式時，27位元（=3 + 24) 16色模式時，7位元（=3 + 4) 因之，整個畫面所需之量爲： 256 X 120 X ( 27 + 7) = 1，〇44,480 位元 =130,560位元組 =127.5K位元組如與只有16M色模式一種者比較，已有改進，但由於色模式愈多，標誌碼之尺寸也要愈多，仍難以期待有多大效果。此例子雖稍偏於極端，但仍表示若不用相當大程度的群加以統合，則有增大標誌碼之危險性。於是，利用電視畫面顯不之方向爲水平方向之事實，乃以16H單位施行模式變換。而且使16M色模式之自然畫像能與128色，64色，32色，16色模式混合存在。在圖3 3之例子中，畫面係明確分成兩半，但如圖3 5所示具有凹凸者，若在16H內混含16色模式及16M色模式兩種時，則設定於16M色模式。如此即可不致減低記憶效率，又能正確再現圖像。雖把畫面分割爲16H單位，但考慮記憶效率，如同一色模式相連時，則把先前之色模式照原引接。依此種方式時，色模式之設定只需2處即可。此時，假定標誌碼爲8 • 26- ^紙張尺度通用中國國家標準(CNS )八4*見格Y2T0X297公廣) "" * (請先閲讀背面之注意事項再填寫本頁) *va 'l· k Α7 、----Β7__ 五、發明説明（24 ) 位元時，顯示圖3 3之圖像所需要之資料尺寸爲： 256 X 120 X (4 + 24) + 8 + 8 = 860，176 位元 =107,522位元組 =約105Κ位元組此値僅爲以16Μ色模式單種處理時之58¾。圖3 6表示其關係。依本發明，在處理混含自然畫像或人造動畫像之圖像時，可增加記憶體之利用效率，亦可大幅提高資料傳輸速度。依習知技術，解決解碼誤差之對策係在傳輸資料中插入再起動標誌碼，而與此相同之效果亦見於16H單位之區塊化中。此乃因由於以16H單位決定色模式，從而可獲知逸離此模式之色碼之資料就是誤差之故》圖式之簡單說明圖1爲Η遮沒及V遮沒之說明圖。圖2爲色向量與點之關係的說明圖。圖3爲虛擬畫面與實畫面之關係之說明圖。圖4爲畫面捲動之說明圖。圖5爲點與調色盤之關係的說明圖。圖6爲可處理自然畫像之圖像處理系統之一例的系統方塊圖。圖7爲圖像資料擴展單元之方塊圖。圖8爲圖像資料之流動與其時序的槪念說明圖。 -27- 本紙張尺度適用中國國家標準（CNS) A4規格（2丨0><297公廣） (請先閲讀背面之注意事項再填寫本頁) -* 經濟部智慧財產局員工消費合作社印製 A 7 ______B7__ 25 五、發明説明（）圖9爲圖像資料擴展單元之暫存器之說明圖。圖1 〇爲16M色模式時之向量因子與點之關係的說明圖。圖1 1爲調色盤之槪念說明圖。圖12爲圖像資料之傳送說明圖。圖13爲壓縮資料之形式與畫面之關係的說明圖。圖1 4爲以4位元固定行進長度，調色盤碼爲可變長度時之記錄長度之說明圖。圖1 5爲表示與圖4之資料形式對應之壓縮資料之解碼情況。圖16表示與圖4資料形式對應之壓縮資料之解碼情況。圖1 7爲單色（2色）模式時與16色模時之顯示色的比較用之圖像例子。圖18爲本發明中之行進長度之資料格式。圖1 9爲畫面分割，資料傳送之單位之說明圖。圖2 0爲傳送資料之形式之說明圖。圖21爲水平捲動量爲零時之傳送時序說明圖。圖2 2爲水平捲動量爲+ 1點時之傳送時序說明圖。圖2 3爲水平捲動量爲-1點時之傳送時序說明圖。圖2 4~2 6爲水平捲動時向視頻編碼單元送出之時序說明圖。圖2 7爲圖像資料擴展單元/視頻編碼單元介面之說明圖。 -28 - 本紙張尺度適用中國國家標準（CNS ) A4規格（210X297公釐） (請先閲讀背面之注意事項再填寫本頁).OT Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs A7 __B7_ V. Invention Description (23) and the mode change is performed per 1 pixel unit. Because the flag code representing the color mode requires 3 bits, the number of bits necessary to represent a pixel is: 27 bits (= 3 + 24) in 16M color mode, 7 bits (= 3 + 4) Therefore, the amount required for the entire picture is: 256 X 120 X (27 + 7) = 1, 〇44,480 bytes = 130,560 bytes = 127.5K bytes, such as the one with only 16M color mode The comparison has been improved, but the more the color mode, the more the size of the logo code. Although this example is slightly extreme, it still shows the danger of increasing the code if it is not integrated to a considerable degree. Therefore, using the fact that the direction in which the television picture is displayed is horizontal, the mode conversion is performed in units of 16H. In addition, natural images in 16M color mode can be mixed with 128, 64, 32, and 16 color modes. In the example of FIG. 33, the screen is clearly divided into two halves, but those with bumps as shown in FIG. 35 are set to the 16M color mode if the 16H mode and the 16M color mode are mixed in 16H. In this way, the memory efficiency is not reduced, and the image can be reproduced correctly. Although the screen is divided into 16H units, considering memory efficiency, if the same color mode is connected, the previous color mode is connected as it is. In this way, only two settings of the color mode are needed. At this time, it is assumed that the mark code is 8 • 26- ^ Common Chinese National Standards (CNS) for paper size 8 4 * See grid Y2T0X297 public broadcast) " " * (Please read the precautions on the back before filling this page) * va 'l · k Α7 、 Β7__ V. Description of the invention (24) The size of the data required to display the image in Figure 33 is: 256 X 120 X (4 + 24) + 8 + 8 = 860,176 bits = 107,522 bytes = approximately 105K bytes This is only 58¾ when single processing in 16M color mode. Figure 36 shows the relationship. According to the present invention, when processing images containing natural portraits or artificial animation images, the utilization efficiency of the memory can be increased, and the data transmission speed can be greatly improved. According to the conventional technology, the countermeasure for solving the decoding error is to insert a restart flag in the transmission data, and the same effect is also seen in the block of 16H units. This is because the color mode is determined in units of 16H, so it can be known that the data of the color code that escapes this mode is the error. "Simple illustration of the figure Figure 1 is an explanatory diagram of Η mask and V mask. FIG. 2 is an explanatory diagram of the relationship between color vectors and dots. FIG. 3 is an explanatory diagram of the relationship between a virtual screen and a real screen. FIG. 4 is an explanatory diagram of screen scrolling. FIG. 5 is an explanatory diagram of a relationship between a dot and a color palette. Fig. 6 is a system block diagram of an example of an image processing system capable of processing natural portraits. FIG. 7 is a block diagram of an image data expansion unit. FIG. 8 is an explanatory diagram of the flow of image data and its timing. -27- This paper size applies to China National Standard (CNS) A4 specifications (2 丨 0 > < 297 public broadcasting) (Please read the precautions on the back before filling out this page)-* Printed by the Employees' Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs System A 7 ______B7__ 25 V. Description of the invention () Figure 9 is an explanatory diagram of the temporary register of the image data expansion unit. Fig. 10 is an explanatory diagram of the relationship between the vector factor and the dots in the 16M color mode. FIG. 11 is an explanatory diagram of the concept of the color palette. FIG. 12 is an explanatory diagram of image data transmission. FIG. 13 is an explanatory diagram of the relationship between the format of compressed data and the screen. Fig. 14 is an explanatory diagram of the recording length when the travel length is fixed at 4 bits and the palette code is variable length. Figure 15 shows the decoding of the compressed data corresponding to the data format of Figure 4. Fig. 16 shows the decoding of the compressed data corresponding to the data format of Fig. 4. Fig. 17 is an example of an image for comparison of the display colors in the monochrome (2-color) mode and the 16-color mode. FIG. 18 is a data format of a travel length in the present invention. Figure 19 is an explanatory diagram of the screen division and data transmission unit. FIG. 20 is an explanatory diagram of a form of transmitting data. FIG. 21 is a transmission timing explanatory diagram when the horizontal scroll amount is zero. Figure 2 2 is a transmission timing diagram when the horizontal scroll amount is + 1 point. Figure 23 is a transmission timing diagram when the horizontal scrolling amount is -1 points. Figures 2 to 26 are timing diagrams for sending to the video coding unit during horizontal scrolling. Figure 27 is an explanatory diagram of the interface of the image data expansion unit / video coding unit. -28-This paper size applies to Chinese National Standard (CNS) A4 (210X297 mm) (Please read the precautions on the back before filling this page)

、1T 經濟部智慧財產局員工消費合作社印製經濟部中央梯準局貝工消费合作社印装 A7 __B7五、發明説明（26 ) 圖2 8表示水平捲動量與讀入時序。圖2 9表示記憶體上之YYUV之配置。圖3 0爲本發明之實施例，表示控制單元之方塊圖。圖3 1爲壓縮圖像資料格式。圖3 2爲依各色模式而分之壓縮圖像資料格式。圖3 3爲由人造動畫像（BGO)及自然圖像（BGI)合成之圖像之例子。圖3 4爲圖6之圖像之合成的說明圖。圖3 5爲在16H內有不同色模式混合存在時之色模式設定例子。圖3 6爲色模式與資料形式之說明圖。 (請先閲讀背面之注f項再填寫本頁) 訂 •2 本紙浪尺度速用中國國家梂準（CNS ) A4规格（210X297公釐）， 1T Printed by the Consumer Cooperative of Intellectual Property Bureau of the Ministry of Economic Affairs Printed by the Shell Consumer Cooperative of the Central Ladder Bureau of the Ministry of Economic Affairs A7 __B7 V. Description of the Invention (26) Figure 2 8 shows the horizontal scrolling amount and reading sequence. Figure 29 shows the YYUV configuration on the memory. FIG. 30 is a block diagram of a control unit according to an embodiment of the present invention. Figure 31 shows the compressed image data format. Figure 32 shows the compressed image data format according to each color mode. Figure 33 is an example of an image synthesized from artificial animation images (BGO) and natural images (BGI). FIG. 34 is an explanatory diagram of the composition of the image of FIG. 6. FIG. Figure 35 is an example of color mode setting when different color modes are mixed in 16H. Figure 36 is an explanatory diagram of the color mode and data format. (Please read the note f on the back before filling out this page) Order • 2 This paper is a fast-moving China National Standard (CNS) A4 specification (210X297 mm)

Claims

A8 B8 C8 D8 6. Application for patent scope 1. An image processing system, which is characterized by a fixed-length data format in which the unit of color data and travel length is an integer multiple of the length of the transmitted data, and contains such fixed length The format travel length is the data format of the extended format that specifies the travel length for subsequent records at the specified time to perform image processing; and includes dividing the screen into horizontal masking units, and specifying a color mode with each masking unit. Means of transmitting data to handle those who mix natural and dynamic images on the screen. (Please read the notes on the back before filling out this page) Order Printed by the Central Standards Bureau of the Ministry of Economic Affairs and printed by wc Industry and Consumer Cooperatives -30- This paper size applies to China National Standard (CNS) Α4 size (210 × 297 mm)